Matthijs Hollemans

This happens often enough that I wrote a blog post about it: http://machinethink.net/blog/help-core-ml-gives-wrong-output/

This *might* be caused by the fact that the GPU in iOS devices using 16-bit floating point instead of the 32-bit floating point used by Caffe. I would try running...

Yes, simulator uses the CPU not the GPU.

The accuracy for MobileNet is pretty much the same as VGG16. But it's 30 times smaller and about 10 times faster. So I'd definitely use MobileNet over VGG16.

I converted the VOC one. Note that in the TensorFlow models repo is a version of SSD that runs on MobileNets. This is roughly as accurate as Tiny YOLO but...

I haven't converted MobileNet+SSD to Core ML, so there may be issues I don't know about but one issue is that the model is in TensorFlow format so you have...

You don't need to make another model. It should be sufficient to just create a new output on the SSD model, then make an output with the same name in...

Nice! Are you running this on an actual device? I don't have an iOS 11 compatible device just yet so I only ran it in the simulator. But I was...

@austingg There is no way to convert TensorFlow models to Core ML at the moment (only Keras models). In addition, the mlmodel format does not support depthwise convolution, so even...

@gwangsc I agree. Please file a feature request at https://bugreport.apple.com -- that's the only way Apple will listen...